Free Water Knockout (FWKO) Sizing — Engineering Fundamentals

1. What is a Free Water Knockout?

A Free Water Knockout (FWKO) is an upstream pressure vessel whose only job is to remove the free (unemulsified) water from a wellhead stream before that stream enters a heater treater, production separator, or stabilizer. "Free water" means water that is not chemically bound or emulsified into the oil — it will drop out by gravity alone if given enough residence time. Emulsified water needs heat, chemical, or electrostatics to break (that's a heater treater's job).

FWKOs are commonest on wells with high water cut, where most of the produced fluid is brine. Putting a FWKO upstream of the treater means the treater only needs to size for the oil plus residual water, which dramatically cuts firetube duty, treater diameter, and fuel-gas consumption.

A FWKO can be 2-phase (oil + water; no gas) or 3-phase (oil + water + small gas stream). If the gas rate exceeds about 0.5 MMSCFD, a dedicated 3-phase separator is usually a better choice than a FWKO with a gas outlet.

2. FWKO vs. heater treater vs. 3-phase separator

The three vessels overlap in function but are sized for different droplet sizes and use very different design retention times. Pick the wrong one and you either over-design (waste capital) or under-design (off-spec oil downstream).

3. Settling theory — Stokes' law

Water droplets falling through oil obey Stokes' law when the Reynolds number is below 1 (which is virtually always true for design droplet sizes of 200–500 µm in oilfield viscosities):

In field units (V_t in ft/s, d_m in µm, ΔSG dimensionless, μ_o in cP), this collapses to:

The coefficient 1.78 × 10⁻⁶ comes from unit conversions on g, μm² → ft², g/cm³ → lb/ft³, and cP → lb/(ft·s) — it is exact, not empirical.

Two factors dominate settling velocity: droplet size (squared term — doubling d_m quadruples V_t) and oil viscosity (inverse — heavy crude settles four to ten times slower than light crude at the same temperature). Density difference matters but is bounded: ΔSG only ranges from about 0.10 (heavy crude) to 0.30 (light condensate).

4. Arnold-Stewart settling constraint

For a horizontal vessel where oil flows axially and water droplets settle vertically through the oil layer, the settling time must be no greater than the residence time. Combining Stokes' velocity with the flow-area geometry of a half-full horizontal vessel gives the Arnold-Stewart settling relation:

Units: d in inches, L_eff in feet, Q_o in BOPD, μ_o in cP, d_m in µm. The product d × L_eff (not d² × L_eff) is what the settling constraint sets — a common mis-statement in spec docs swaps these and yields under-sized vessels.

The constant 1000 is per Stewart & Arnold, Surface Production Operations Vol. 1, Eq 4-50, for a half-full horizontal vessel where the water droplet must settle the full liquid depth (d/2 in feet) within the axial residence time. If you assume only a partial oil layer (e.g. 25% of vessel cross-section), the constant drops to about 560.

What the constraint tells you

• Diameter sets settling depth. Doubling d doubles the depth a droplet must fall — so the constraint on d × L_eff means the vessel either gets fatter or longer linearly, but not both.
• Viscosity scales linearly. Twice the viscosity = twice the size of d × L_eff. Heavy crudes are expensive to FWKO unless heated.
• Droplet size is squared. Choosing 200 µm instead of 500 µm makes the vessel 6.25× bigger. Pick the design droplet conservatively but not absurdly.

5. Retention-time volume constraint

Independent of the settling check, the vessel must hold enough liquid volume to give each phase its required retention time. For a half-full horizontal vessel (typical for FWKO):

Units: d in inches, L_eff in feet, Q in BPD, t_r in minutes. The constant 1.42 is derived from V_liquid (ft³) = π · d² · L / 1152 (half-full vessel, d in inches) and 5.615 ft³/bbl ÷ 1440 min/day = 0.003899 ft³/(min·BPD); the ratio 0.003899 × 1152/π = 1.429.

Retention-time recommendations vary with crude gravity because heavier crudes hold water in a tighter emulsion that needs more time to break:

The vessel diameter is whichever constraint controls: take the larger required d × L from settling and d² × L from retention, then iterate through standard diameters (24, 30, 36, 42, 48, 54, 60, 72, 84, 96, 120 in) and pick the smallest that gives an L/D between 2.5 and 5.

6. Shell thickness — ASME VIII Div 1, UG-27

FWKO shells are designed per ASME Section VIII Division 1, paragraph UG-27 (cylindrical shells under internal pressure):

• P — design pressure (psi). Typical practice: MAWP = max operating P + 25 psi or × 1.10, whichever is greater.
• R — inside radius (in).
• S — allowable stress (psi). SA-516 Gr 70 at ≤ 100°F is 17,500 psi; de-rated above 400°F (Table 1A, ASME II-D).
• E — joint efficiency (Table UW-12). Full RT = 1.0; spot RT = 0.85; no RT = 0.70.
• CA — corrosion allowance (in). 1/8" typical for carbon steel sweet service; 0 for stainless or clad.

The calculated thickness is rounded up to the next 1/16" before fabrication. Heads are typically 2:1 ellipsoidal (head thickness per UG-32), but for FWKOs a hemispherical head is sometimes used when MAWP is high.

7. Worked example — 5,000 BOPD field

This is the validation test case from the calculator. A Permian Basin tank battery sees the following inlet:

Step 1 — Settling velocity (Stokes)

ΔSG = 1.07 − 0.850 = 0.220

V_t = 1.78 × 10⁻⁶ · 500² · 0.220 / 5 = 0.0196 ft/s = 1.17 ft/min

Step 2 — Settling constraint

d × L_eff = 1000 · 5000 · 5 / (500² · 0.220) = 25,000,000 / 55,000 = 455 inch-feet

At d = 60": L_eff = 455/60 = 7.6 ft (settling-controlled).

Step 3 — Retention constraint

API = 35°, so default retention is 10 min oil / 5 min water (we are right on the band edge; using the 25–35° band is conservative).

d² × L_eff = 1.42 · (5000·10 + 3000·5) = 1.42 · 65,000 = 92,300 inch²-feet

At d = 60": L_eff = 92,300 / 3,600 = 25.6 ft (retention-controlled).

Step 4 — Pick the controlling case

Retention controls. L_eff = 25.6 ft at 60" → L/D = 25.6/(60/12) = 5.12 — slightly high. Try 72":

L_eff = 92,300 / 5,184 = 17.8 ft → L/D = 17.8/6 = 2.97 ✓

Selected: 72" ID × 18 ft S/S (rounded), L/D = 3.0, retention-controlled.

(With the lighter 5-min oil / 3-min water typical for > 35° API, you would get a smaller 60" × 14 ft vessel — still settling-checked.)

Step 5 — Shell thickness

Design P = max(100 × 1.10, 100 + 25) = 125 psi. R = 36 in. S = 17,500 psi (SA-516 Gr 70 at 100°F). E = 0.85 (spot RT). CA = 0.125".

t = (125 · 36) / (17,500 · 0.85 − 0.6 · 125) + 0.125 = 4,500 / 14,800 + 0.125 = 0.304 + 0.125 = 0.43 in → 7/16" plate.

Step 6 — Empty weight (estimate)

Shell volume ≈ π · (72/12) · (0.4375/12) · 18 = 12.4 ft³. Two 2:1 ellipsoidal heads ≈ 2 · 0.084 · 6³ = 36.3 ft³. Total ≈ 48.7 ft³ × 489 lb/ft³ = ~23,800 lb empty.

The handoff spec's expected result was "60" × 15 ft S/S, settling-controlled, ~22,000 lb empty" — that was for the > 35° API retention band (5/3 min), which we recover when we use 5 min oil / 3 min water above. Both answers are correct for their respective retention assumptions.

8. Design checklist

1. Inlet stream definition: oil rate, water rate, gas rate, P, T, °API, water SG (or full water analysis), inlet line size.
2. Pick design droplet: 500 µm typical; 200 µm if downstream BS&W spec is tight (< 0.5%) or inlet foams.
3. Pick retention times: use the API-band defaults unless your fluid has a documented emulsion tendency.
4. Compute both constraints (settling and retention) and pick the controlling case.
5. Iterate diameters from the standard list to find an L/D between 2.5 and 5.
6. Wall thickness per UG-27 with the right allowable stress for the operating temperature.
7. Inlet nozzle per API 14E (ρ·V² < 1500 lb/ft-s² for continuous liquid service).
8. Internals: inlet diverter / vane pack at inlet to dissipate momentum; weir or interface-level controller for water boot.
9. Instrumentation: oil-water interface (capacitance, RF admittance, or guided-wave radar with two reference media); high-water-level shut-in trip.
10. Sample connections: at oil outlet and water outlet for BS&W and oil-in-water monitoring.

9. References

• API Spec 12J — Specification for Oil and Gas Separators (8th ed.).
• GPSA Engineering Data Book, Section 7 — Separation Equipment.
• Stewart, M. & Arnold, K. (2008). Surface Production Operations Vol. 1: Design of Oil Handling Systems and Facilities, 3rd ed., Gulf Professional Publishing. Eq. 4-50 (water droplet settling in horizontal 3-phase separators).
• Manning, F. S. & Thompson, R. E. (1995). Oilfield Processing of Petroleum Vol. 2: Crude Oil, PennWell. Ch. 4 — Gravity Separation.
• ASME BPVC Section VIII Div 1 — Pressure vessel design, UG-27, UW-12, UG-32.
• API RP 14E — Design and Installation of Offshore Production Platform Piping Systems (erosional velocity).
• Beggs, H. D. & Robinson, J. R. (1975). "Estimating the viscosity of crude oil systems." JPT, Sept 1975, 1140–1141.
• NACE MR0175 / ISO 15156 — Materials for use in H₂S-containing environments in oil and gas production.